DataverseNO
Not a member yet
2167 research outputs found
Sort by
Replication Data for: L'acquisition des voyelles nasales en français : une étude acoustique et perceptive sur la prononciation des apprenants néerlandophones belges
No full text Dataset abstract
This dataset contains two types of data on the production accuracy of French nasal vowels realized by L1 Belgian Dutch learners, i.e. listener-based and acoustic measures. By focusing on these two measures, we shed light on two different dimensions of production accuracy, i.e. vowel intelligibility and phonetic nativelikeness. First, this dataset contains acoustic data of 20 L1 Belgian Dutch speakers and 12 L1 Northern Metropolitan French speakers (Carignan, 2014). Vowels were produced in high-frequency monosyllabic French words during a reading task. F1 and F2 values were calculated for the midpoint of each vowel and normalized using Lobanov z-score calculation both across and within speaker groups. Secondly, this dataset contains perceptual data of 71 L1 French speakers respectively representing the Ile-de-France region (France) and Liège (Belgium). Participants performed an online identification task that assessed both listeners’ actual understanding of non-native accented nasal vowels and their category goodness judgments on a 5-point scale (1 = "bad", 5 = "good"). Article abstract
This paper examines the production of French nasal vowels by Belgian Dutch learners. It is innovative in that it combines the analysis of ‘intelligibility’ of non-native accented vowels, assessed by listeners representing Paris and Liège, with an acoustic analysis focusing on ‘phonetic nativeness’. The stimuli were high-frequency monosyllabic words produced by 20 Belgian Dutch learners. The results indicate that the intelligibility of non-native accented nasals is rather high, except for the /ɑ̃/-vowel. Furthermore, the two listener groups do not differ in their actual understanding of vowels, except for the /ɛ̃/-vowel, which is identified more often by listeners representing the Liège region. This difference may be related to the impact of (i) L1 variation or (ii) Belgian listeners’ familiarity with Dutch-accented speech on speech perception. Finally, the study shows that it is not so much nasality itself that poses production difficulties for Dutch-speaking learners, but the interaction with vowel quality.</p
Replication Data for: The 3-billion fossil question: How to automate classification of microfossils
No full textThis dataset consists of 100,000 PNG images of individual microfossils extracted from whole slide images of fossils from a single wellbore. The well in question belongs to the Mikkel field on the Norwegian continental shelf. The wellbore is located at 64 degrees north-south, 7 degrees east-west with ID NO 6407-6-5. All images in this dataset are of RGB format with a 224-by-224 pixel resolution. There are no labels associated with these images, thus the number of different species represented in this dataset is unknown and likely to be in the order of 1000. The dataset was created using the method described in the associated paper. The purpose was to create a medium sized dataset of preprocessed microfossil crops for use in self-supervised training of small to medium sized deep learning models, for which this dataset is suffiently big. For large scale training of e.g. Vision Transformers, more data will be required. See the following fact page for more geological information: https://factpages.sodir.no/en/wellbore/PageView/Exploration/With/PalySlides/3921. Note that the original whole slide images from which this dataset is created were provided by the Norwegian national data repository for petroleum data (Diskos) under the Norwegian Licence for Open Government Data (NLOD) 2.0
Replication data for: Near-surface characterization using shear-wave resonances: A case study from offshore Svalbard, Norway
No full textThis data set is used to conduct research as presented in the manuscript entitled "Near-surface characterization using shear-wave resonances: A case study from offshore Svalbard, Norway" written by Taweesintananon et al. in 2024, which is submitted for peer reviews and publications in a journal. It contains processed strain data in nanostrain unit measured by OptoDAS interrogator through the dark fibers in two submarine telecommunication cables, which were installed into soft sediments at 0–2 m below the seafloor, between Longyearbyen and Ny-Ålesund in Svalbard, Norway. In total, four OptoDAS interrogators are used for DAS data recording: the ends of the inner (L2) and outer (L3) cables in Longyearbyen, and the other ends of the inner (N2) and outer (N3) cables in Ny-Ålesund. The raw data were acquired in August, 2022, and converted to strain data in nanostrain unit. To reduce file sizes, we resampled the DAS strain data from 1.6 to 20 ms time sampling intervals with an antialiasing filter at 80% of the output Nyquist frequency of 25 Hz. After that, we attenuated interrogator noise that occurs in every DAS recording channel. First, we obtained the single-trace interrogator noise model by stacking (a mean calculation) all the DAS data traces in the first 40 km of the cable, in which the instrument noise level was low. Then, we subtracted every trace of the DAS data by the interrogator noise model. The resampled data after the interrogator noise removal have been archived and used in our spectral analysis. The processed data and the power spectra are stored as structured arrays in the MATLAB file format. This data set has no exact positioning data of the fiber optic cable, because they are proprietary to Sikt. However, the distance of each recording channel along the cable from the shore is given in this data set
Supporting Data for: Bilayer Membranes with Frequent Flip-Flops Have Tensionless Leaflets
No full textDataset contains long MARTINI (v 2.2) simulations of GM1-containing lipid membranes of POPC (POPC = 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine, GM1 = ganglioside GM1 or monosialotetrahexosylganglioside). Only one monolayer contains GM1s and the opposite monolayer contains the amount of POPC, that makes surface tension of both monolayers almost equal (see Fig. 3 in the reference paper).
Every folder contains the archive "misc.zip", which in turn contains everything for reproduction of the simulations. Please use TREE VIEW to browse files efficiently
Replication Data for the Dataset of the Histories of Reception of Photography in the Ottoman Empire project
No full textThis dataset contains the data collected for the Histories of Reception of Photography in the Ottoman Empire project, undertaken as a Marie Skłodowska-Curie Action project at the Department of Philosophy, Classics, History of Art and Ideas (IFIKK) at the University of Oslo (UiO) from 01.09.2022 until 31.08.2024. The dataset consists of the books and articles on photography written in Western Armenian and in Ottoman Turkish in the late 19th and early 20th century
Replication Data for: Learning to predict - second language perception of reduced multi-word sequences
No full textDATASET ABSTRACT
This is the data and code from a word-monitoring task, in which advanced learners of English responded to the word 'to' in verb + to-infinitive structures (V-to-Vinf) in English, where 'to' could occur in a full or reduced pronunciation (e.g. "prefer to" [tʊ] or "preferda" [ɾə]). The design of this experiment is replicated from our earlier study with American English native speakers (Lorenz & Tizón-Couto, 2019, see link to paper and dataset below *).
We tested the effects of string frequency (V+to) and transitional probability (of 'to' given the V) on the accuracy and speed of recognition of "to" in spoken sentences. These effects were analysed with mixed-effects generalized additive models (GAMM); the code also includes visualisations of these models.
The experiment was run with OpenSesame (version 3.2.6 for Mac, see Mathôt et al. 2012). The data include information on frequencies of occurrence of words and bigrams; this was extracted from the Corpus of Contemporary American English (COCA, Davies 2008–). We used R (version 4.3.1, R Core Team 2023) for all data analyses, hence the code can best be replicated in R.
*) Lorenz, D. & Tizón-Couto, D. (2019). Chunking or predicting – frequency information and reduction in the perception of multi-word sequences. Cognitive Linguistics, 30(4), 751-784.
https://doi.org/10.1515/cog-2017-0138 (the paper); https://doi.org/10.18710/7TSABU (the data)
PUBLICATION ABSTRACT
The cognitive entrenchment of frequent sequences comes as ‘chunking’ (holistic storage) and as ‘procedure strengthening’ (predicting elements in a sequence). A growing body of research shows effects of entrenchment of multi-word sequences in the native language, which is learned and shaped continuously and intuitively. But how do they affect L2 speakers, whose language acquisition is more analytic but who nonetheless also learn through usage?
The present study tests advanced English learners’ receptive processing of multi-word sequences with a word-monitoring experiment. Recognition of to in the construction V-to-Vinf was tested for full and reduced forms ([tʊ] vs [ɾə]), conditioned by the general frequency of the V-to sequence and the transitional probability (TP) of to given the verb (V > to). The results are compared with those previously obtained from native speakers (Lorenz & Tizón-Couto, 2019).
Results show that recognition profits from surface frequency, but not from TP. Reduced forms delay recognition, but this is mitigated in high- frequency sequences. Unlike native speakers, advanced learners do not exhibit a chunking effect of high-frequency reduced forms, and no facilitating effect of TP.
We attribute these findings to learners’ lesser experience with spontaneous speech and phonetic reduction. They recognize reduced forms less easily, show weaker entrenchment of holistic representations, and do not draw on the full range of probabilistic cues available to native speakers.</p
Replication Data for: Exploring multimorbidity patterns in older hospitalized Norwegian patients using network analysis modularity
No full text The repository contains two datset files. An edge list neede to replicate the network multimorbidity modules. The edge list is a data structure used to represent a graph as a list of its edges. And a dataset for multimorbidity patterns (modules) with modules' diseases description. The datasets serve as a supplementary material for the publication "Exploring multimorbidity patterns in older hospitalized Norwegian patients using network analysis modularity".
The aim of this study is to demonstrate the use of Network Analysis (NA) in describing the Multimorbidity Patterns (Mps) in the Norwegian hospitalized older patient population based on national register data.
The source of the original data used in the project is the Norwegian Patient Register (NPR) which contains administrative, demographic, and medical data of patients who have been in contact with specialized health services [16].
Our data cohort comprises all admissions of the patients who were admitted to a Norwegian hospital in a span of three years (2017-2019).</p
Oceanographic time series at South Cape from April 2022 to April 2023
No full textTime series of temperature, salinity, pressure, pH and Oxygen obtained from a Seabird SeapHOx sensor deployed south of Svalbard to monitor an intense seepage. Conservative temperature, potential temperature and absolute salinity were calculated and added to the dataset
Replication Data for: Understanding ‘many’ through the lens of Ukrainian багато
No full textDataset description:
The General Regionally Annotated Corpus of Ukrainian (GRAC, Shvedova et al. 2017-2024, uacorpus.org) was consulted to collect data for further analysis concerning the distribution of Singular vs. Plural verb forms in the target bahato construction. GRAC is a Sketch Engine corpus of over 1.8 billion words, representing texts from over 30,000 authors created between 1816 and 2023. This corpus is designed to serve as source material for linguistic research on Standard Ukrainian. Our data was collected during the month of February 2024. We extracted and annotated 28,491 examples of the bahato construction.
An additional set of examples was collected from the Russian National Corpus (ruscorpora.ru) during the month of August 2024 to provide comparison with the Russian mnogo construction. For this purpose, 6,612 examples were extracted and annotated for word order and Singular vs. Plural verb agreement.
Both the Ukrainian and the Russian data are included in this dataset, along with the R scripts used to analyze this data. Article abstract:
We reveal an ongoing language change in Ukrainian involving a construction with a subject comprised of the indefinite quantifier багато ‘many’ modifying a noun phrase in the Genitive Plural. Number agreement on the verb varies, allowing both Singular (in 69.1% of attestations) and Plural (in 30.9% of attestations). Based on statistical analysis of corpus data, we investigate the influence of the factors of year of creation, word order of subject and verb, and animacy of the subject on the choice of verb number. We find that, while all combinations of word order and animacy are robustly attested, VS word order and inanimate subjects tend to prefer Singular, whereas SV word order and animate subjects tend to prefer Plural. Since about the 1950s, the proportion of Plural has been increasing, overtaking Singular in the current decade. We propose that this Singular vs. Plural variation is motivated by the human embodied experience of construing a group of items as either a homogeneous mass (and therefore Singular) or a multiplicity of individuals (and therefore Plural). This proposal is supported by the identification of micro-constructions that prefer Singular and show reduced individuation of human beings
Replication Data for: A numerical study of the R744 primary cooling system for ATLAS and CMS LHC detectors
No full textA new R744 refrigeration system is needed to cool down the detectors, namely ATLAS and CMS, of the Large Hadron Collider (LHC) at CERN in Geneva. A R744-R744 cascade system is planned having a primary 2-stage compression loop and an oil-free, pumped secondary loop. To guarantee a stable and reliable operation of the plant, on one hand, two demonstration plants are built: System A for the evaluation of its control logic and System B for the analysis of the behavior of the refrigerant regarding the height difference (78m). On the other hand, simulations are done in Dymola tool.
Two data sets extracted from the demonstration plant (System A) on-site as well as the results of the simulations are stored here